FAULT DETECTION AND DIAGNOSIS METHOD FOR THREE-PHASE

INDUCTION MOTOR

ROZBEH YOUSEFI

A thesis submitted in fulfillment of the

requirements for the award of the degree of

Doctor of Philosophy (Electrical Engineering)

School of Electrical Engineering

Faculty of Engineering

Universiti Teknologi Malaysia

NOVEMBER 2018

iii

To my beloved wife, Elham

To my beloved father and mother

To my beloved brother

iv

ACKNOWLEDGEMENT

I would like to thank my principal supervisor, Prof. Datin Dr. Rubiyah bt. Yusof

for her guidance during my research and study. Her perpetual energy and enthusiasm in

research had motivated me.

v

ABSTRACT

Induction motors (IM) are critical components in many industrial processes. There

is a continually increasing interest in the IMs’ fault diagnosis. The scope of this thesis

involves condition monitoring and fault detection of three phase IMs. Different

monitoring techniques have been used for fault detection on IMs. Vibration and stator

current monitoring have gained privilege in literature and in the industry for fault

diagnosis. The performance of the vibration and stator current setups was compared and

evaluated. In that perspective, a number of data were captured from different faulty and

healthy IMs by vibration and current sensors. The Principal Component Analysis (PCA)

was utilized for feature extraction to monitor and classify collected data for finding the

faults in IMs. A new method was proposed with the combined use of vibration and current

setups for fault detection. It consists of two steps: firstly, the training part with the aim of

giving acceleration property (nature of vibration data) to the current features, and secondly

the testing part with the aim of excluding the vibration setup from the fault detection

algorithm, while the output data have the property of vibration features. The 0-1 loss

function was applied to show the accuracy of vibration, current and proposed fault

detection method. The PCA classified results showed mixed and unseparated features for

the current setup. The vibration setup and the proposed method resulted in substantial

classified features. The 0-1 loss function results showed that the vibration setup and the

developed method can provide a good level of accuracy. The vibration setup attained the

highest accuracy of 98.2% in training and 92% in testing. The proposed method performed

well with accuracies of 96.5% in training and 84% in testing. The current setup, however,

attained the lowest level of accuracy (66.7% in training and 52% in testing). To assess the

performance of the proposed method, the Confusion matrix of classification in NN was

utilized. The Confusion matrix showed an accuracy of 95.1% of accuracy and negligible

incorrect responses (4.9%), meaning that the proposed fault detection method is reliable

with minimum possible errors. These vibration, current and proposed fault detection

methods were also evaluated in terms of cost. The proposed method provided an

affordable fault detection technique with a high accuracy applicable in various industrial

fields.

vi

ABSTRAK

Induction motor (IM) adalah komponen kritikal dalam banyak proses

perindustrian. Terdapat minat yang semakin meningkat dalam diagnosis IMs. Skop tesis

ini melibatkan pemantauan keadaan dan pengesanan kesalahan tiga fasa IMs.Teknik

pemantauan yang berbeza telah digunakan untuk pengesanan kesalahan pada IM. Getaran

dan stator pemantauan semasa telah mendapat keistimewaan dalam banyak kajian dan

dalam industri untuk diagnosis kesilapan. Prestasi getaran dan tetapan semasa stator telah

dibandingkan dan dinilai. Dalam perspektif itu, beberapadata telah diambil dari pelbagai

IM yang elok dengan getaran dan penderia semasa. Analisis Komponen Utama (PCA)

digunakan untuk pengekstrakan ciri untuk memantau dan mengklasifikasikan data yang

dikumpul untuk mencari kesalahan dalam IM. Kaedah baru dicadangkan menggunakan

gabungan getaran dan tetapan semasa untuk pengesanan kesalahan terdiri daripada dua

langkah: bahagian latihan dengan tujuan memberikan pecutan harta (sifat data getaran)

kepada ciri-ciri semasa, dan sebahagian ujian dengan tujuan pengecualian persedian

ediaan getaran dari algoritma pengesanan kesalahan, sementara data output mempunyai

sifat ciri getaran. Fungsi kerugian 0-1 digunakan untuk menunjukkan ketepatan getaran,

kaedah pengesanan kesalahan semasa dan cadangan yang dicadangkan. Hasil

pengklasifikasian PCA menunjukkan ciri bercampur dan tidak terpakai untuk persediaan

semasa. Persediaan getaran dan kaedah yang dicadangkan menghasilkan ciri-ciri terkelas

yang besar. Hasil fungsi kehilangan 0-1 menunjukkan bahawa persediaan getaran dan

kaedah yang dibangunkan dapat memberikan ketepatan yang baik. Persediaan getaran

mengakibatkan ketepatan tertinggi 98.2% dalam latihan dan 92% dalam ujian. Kaedah

yang dicadangkan dijalankan dengan baik dengan ketepatan 96.5% dalam latihan dan 84%

dalam ujian. Walau bagaimanapun, persediaan semasa mengakibatkan tahap ketepatan

minimum (66.7% dalam latihan dan 52% dalam ujian). Untuk menilai prestasi kaedah

yang dicadangkan, klasifikasi kekeliruan klasifikasi dalam NN digunakan. Matriks

kekeliruan menunjukkan 95.1% ketepatan dan tindak balas yang tidak dapat diabaikan

(4.9%), yang bermaksud bahawa kaedah pengesanan kesalahan yang dicadangkan boleh

dipercayai dengan ralat minimum yang mungkin. Kaedah getaran, semasa dan cadangan

pengesanan kesalahan ini juga dinilai dari segi kos. Kaedah yang dicadangkan

menyediakan teknik pengesanan kesalahan berpatutan dengan ketepatan tinggi yang

digunakan dalam pelbagai bidang perindustrian.

vii

TABLE OF CONTENTS

CHAPTER TITLE PAGE

DECLARATION ii

DEDICATION iii

ACKNOWLEDGEMENTS iv

ABSTRACT v

ABSTRAK vi

TABLE OF CONTENTS vii

LIST OF FIGURES xi

LIST OF TABLES xiv

LIST OF ABBREVIATIONS xv

LIST OF SYMBOLS xvii

LIST OF APPENDICES xviii

1 INTRODUCTION 1

1.1

1.2

Faults in Induction Motors

Maintenance Strategies

1.2.1 Condition Monitoring for Fault Diagnosis

Prediction

1

2

4

1.3 Methods for Fault Diagnosis of IMs

1.3.1 Model-Based Fault Detection Method

1.3.2 Signal-Based Fault Detection Method

1.3.3 Data-Based Fault Detection Method

5

5

6

7

1.4 AI Techniques for Motor Fault Diagnosis 9

1.5

1.6

Problem Statement

Objectives of the Study

10

13

viii

1.7 Scope of the Study 13

1.8 Thesis Organization 14

2 LITERATURE REVIEW 16

2.1 Induction Motor IIM) 16

2.2 Induction Motor faults

2.2.1 Electrical Faults

2.2.1.1 Stator Faults

2.2.1.2 Rotor Faults

2.2.2 Mechanical Faults

2.2.2.1 Bearing Faults

2.2.2.2 Eccentricity Faults

18

19

21

21

22

22

23

2.3

2.4

2.5

2.6

2.7

2.8

Condition Monitoring

Condition Monitoring Techniques

2.4.1 Noise Monitoring

2.4.2 Magnetic Flux Monitoring

2.4.3 Partial Discharge Monitoring

2.4.4 Thermal Monitoring

2.4.5 Air Gap Torque Monitoring

2.4.6 Vibration Monitoring

2.4.7 Stator Current Monitoring

Fault Detection Procedure

2.5.1 Feature Extraction

2.5.1.1 Principal Component Analysis (PCA)

Classification Process

Artificial Intelligence (AI)-based Techniques in

Induction Motor Fault Diagnosis

2.7.1 Artificial Neural Networks (ANNs)

2.7.1.1 Unsupervised Training

2.7.1.2 Supervised Training

Chapter Summary

24

25

26

26

27

27

28

29

30

31

32

33

38

39

41

42

43

45

ix

3 RESEARCH METHODOLOGY 47

3.1

3.2

3.3

3.4

3.5

3.6

Introduction

Fault Detection with Vibration and Current Setups

3.2.1 Classification Using Multilayer NN

3.2.2 0-1 Loss Function

NN Nonlinear Regression

Proposed Method with Joint Use of Vibration and

Current Setups

3.4.1 Relation between Current and Vibration

Experimental Setup

3.5.1 Hardware Setup

3.5.2 Data Acquisition

3.5.3 Training and Testing

Chapter Summary

47

48

52

57

58

61

64

65

65

67

69

69

4 RESULT AND DISCUSSION 71

4.1

4.2

4.3

4.4

4.5

4.6

Introduction

Effectiveness of Vibration and Current Signals for

Fault Diagnosis

Performance Evaluation of Vibration and Current

Setups

Performance of the Proposed Method for Fault

Diagnosis

Cost Effectiveness

Chapter Summary

71

71

76

79

83

87

5

CONCLUSION

5.1 Introduction

5.2 Significant Findings

5.3 Feature Work Recommendations

88

88

88

90

x

REFERENCES 91

Appendices A – C

103

xi

LIST OF FIGURES

FIGURE NO.

TITLE PAGE

1.1

1.2

1.3

1.4

Three main maintenance strategies

Model-based diagnostic technique. Two techniques are

possible according to the same basic structure

Block diagram of signal-based diagnostic procedure

Block diagram of data-based diagnostic procedure

3

5

7

8

2.1 Typical three-phase induction motor

17

2.2 Classification of IM faults

18

2.3

2.4

2.5

Percentage (%) component of IM failure

Burned out stator winding faults

Broken rotor faults

19

20

20

2.6 Different type of eccentricity; (a) without eccentricity; (b)

Static eccentricity; (c) Dynamic eccentricity; (d) Mixed

eccentricity

24

2.7 Block diagram of fault diagnostic procedure

32

2.8 PCA for data representation

34

2.9 PCA for dimension reduction

35

2.10

2.11

2.12

3.1

The PCA transformation

Block diagram of unsupervised training

Block diagram of supervised training

Fault detection procedure with vibration and current setup

38

43

45

48

xii

3.2

3.3

3.4

3.5

3.6

3.7

3.8

3.9

3.10

3.11

3.12

3.13

3.14

4.1

4.2

4.3

4.4

4.5

4.6

Plot of classification accuracy on train and validation

datasets showing an underfit model

Plot of classification accuracy on train and validation

datasets showing an overfit model

Plot of classification accuracy on train and validation

datasets showing a fit model

NN classifier architecture

Graphical ReLU function

Plot of model loss on training and validation datasets

Plot of a fit model loss on training and validation datasets

NN nonlinear regression architecture

Flowchart of proposed methodology

Schematic of healthy three phase induction motor with its

specifications (a), faulty induction motor with stator

winding fault (b) and faulty induction motor with broken

rotor bar fault (c)

Experimental setup for the three-phase induction motor

with accelerometer installed

Data acquisition diagram

Induction motor control block diagram

Data acquisition for vibration and current monitoring at

intervals of 40s

Data acquisition for vibration monitoring

Data acquisition for current monitoring

Gaussion diagram for (a) vibartion and (b) current signals

PCA classified results from the data captured by vibration

sensor

PCA classified results from the data captured by electrical

current sensor

53

54

54

55

56

59

60

60

62

65

66

68

68

72

73

74

75

76

77

xiii

4.7

4.8

4.9

4.10

4.11

Conversion process in the proposed method

Classified results of the data captured by current sensor

using NNs based on the vibration data

Compare the classification results of vibration setup,

current setup and proposed method

Confusion matrix of classification part for proposed method

Break-even point graph

79

80

80

82

85

xv

LIST OF TABLES

TABLE NO.

TITLE PAGE

1.1 Comparison of methods for motor fault detection and

diagnosis

9

3.1

Three-phase Induction motor physical data

66

4.1

4.2

4.3

4.4

4.5

0-1 loss function results for vibration and current setups

Accuracy of current and vibration setups before and after

tapping test

Comparison of vibration setup, current setup and proposed

method with 0-1 loss function

Cost evaluation for current, vibration and proposed fault

detection setup

Advantages and disadvantages of vibration, electrical

current setup and proposed method for fault detection in

IMs

78

78

81

84

86

xv

LIST OF ABBREVIATIONS

AI - Artificial Intelligence

ANN - Artificial Neural Network

AR - Autoregressive

ARMA - Autoregressive Moving Average

CMD - Compact Matrix Decomposition

COR - Correlation

EEG - Electro Encephalogram

ES - Expert System

Expert Systems

FEA - Finite Element Analysis

GA - Genetic Algorithms

ICA - Independent Component Analysis

IM - Induction Motor

KBS - Knowledge-Based Systems

LDA - Linear Discriminate Analysis

LH

- Local Homogeneity

xvi

MA - Moving Average

MSE

- Mean Square Error

MSCA - Motor current Signature Analysis

MMFs - Magneto Motive Forces

NN - Neural Network

PCA - Principal Component Analysis

RMS - Root Mean Square

SCSA - Stator Current-Signature Analysis

SVM - Support Vector Machines

SVD - Singular Value Decomposition

VSI - Voltage Source Inverter

VDI - Verein Deutscher Ingenieure

xvii

LIST OF SYMBOLS

fs - Input stator frequency

Vs - Input stator phase voltage

Ns - Number of stator winding turns

p - Number of pole pairs

Lls - Stator winding leakage inductance

Rs - Stator winding electrical resistance

n - Number of rotor bars

rag - Air-gap average radius

lag - Air-gap length

lr - Rotor length

Lrb - Rotor bar self inductance

Rrb - Rotor bar resistance

xviii

LIST OF APPPENDICES

APPENDIX

TITLE PAGE

A Regression Analysis 103

B Data Acquisition for Machinery Faults Simulator 1 0 4

B1 Data Acquisition for Acceleration Using Labview 104

B2 Creating the Periodic Data Samples 117

C Vibration DAQ Card Quote Form 128

CHAPTER 1

INTRODUCTION

1.1 Faults in Induction Motors

Induction motors (IM) are most commonly used electrical machines in industry

because of their low cost, reasonably small size, ruggedness, low maintenance, and

operation with an easily available power supply (El Hachemi Benbouzid, 2000). However,

they are subjected to different modes of faults leading to failures. These faults can be

inherent to the machine itself or may be created by operating conditions. The inherent

faults could be caused by the mechanical or electrical forces acting on the machine

enclosure. If a fault is not detected or if it is allowed to be developed further, it may lead

to a failure (El Hachemi Benbouzid, 2000; Yamamura, 1979).

The main faults of IMs can generally be classified as 1) stator faults resulting in

the opening or shorting of one or more of the stator phase winding 2) abnormal connection

of the stator windings 3) broken rotor bar or cracked rotor end rings 4) static and dynamic

air gap irregularities 5) bent shaft 6) shorted rotor field winding 7) bearing & gearbox

failures (Bonett & Soukup, 1992; Shashidhara & Raju, 2013; J.-W.Zhang, Zhu, Li, Qi, &

Qing, 2007; Cusidó et al., 2011). The squirrel cage of an induction machine consists of

rotor bars and end rings. Motor with broken bar fault has one or more of the cracked or

broken bars. Broken rotor bar can be caused by manufacturing defects, thermal stresses or

frequent starts of the motor at rated voltage (Siddiqui, Sahay, & Giri, 2014). Winding fault

is due to catastrophe of insulation of the stator winding. This fault can be caused by

mechanical stresses due to movement of stator coil and rotor striking the stator, electrical

2

stresses due to the supply voltage transient or thermal stresses due to thermal overloading

(Siddiqui, Sahay, & Giri, 2014). Whereas faulty IMs must normally be removed from the

application in order to be fixed and repaired, a suitable fault diagnosis and monitoring

system can reduce the financial loss, drastically (Zarei, 2012; P. Zhang, Du, Habetler, &

Lu, 2011).

Early detection of incipient faults is a very important issue in preventive and

predictive maintenance of electrical equipment. Since in modern industries, the majority

of the equipment is driven by three-phase induction motors (IMs), the condition

monitoring of such machines constitutes an essential concern in any industrial section

(Butler, 1996; Cusidó, Romeral, Ortega, Garcia, & Riba, 2011).

1.2 Maintenance Strategies

Traditional machinery maintenance practice in industry can be categorized

broadly into three methods:

a) Run-to-failure maintenance

b) Scheduled maintenance

c) Condition based maintenance

Run-to-failure maintenance, which reacts to the equipment failure after it happens.

This maintenance approach is a corrective management method that has no special

maintenance plan in place. Due to the nature of the industry sectors, the failure of one

piece of equipment may stop production in a significant portion. For example, the failure

of a main haulage belt motor in mining industry may idle an entire mine. In this case, the

run-to-failure maintenance will be too costly. This type of maintenance method is not an

acceptable maintenance method because there might be a high risk of secondary failure,

overtime labor and high cost of spare parts (Palem, 2013; Yam, Tse, Li, & Tu, 2001).

3

Scheduled maintenance, which is the practice of replacing components in fixed

time intervals. This maintenance takes preventive actions to check, repair, or replace the

equipment at a prearranged schedule before machine faults. Such maintenance policy

benefits in terms of maintenance cost reduction as it minimizes the unscheduled downtime

and labor costs in comparison to the run-and-failure maintenance strategy. However, this

strategy does not consider the condition of the equipment in that it scheduled the

maintenance activity at a fixed time interval without considering the condition of the

equipment or component (Palem, 2013; Yam, Tse, Li, & Tu, 2001). In addition, machines

may be repaired when there are no failures (Kwitowski, Lewis, & Mayercheck, 1989).

Condition based monitoring is a maintenance procedure that uses sensors to evaluate the

health of the system. Condition monitoring and fault diagnostics are useful for early

detection of mechanical and electrical failures to prevent main component faults (Jardine,

Lin, & Banjevic, 2006). One of the key elements to condition based maintenance is the

understanding of the actual condition or health of a machine, then using this information

to schedule and perform maintenance when it is most needed. If performed correctly,

condition based maintenance can bring out many advantages such as increasing machinery

availability and performance, reducing consequential damage, increasing machine life,

reducing spare parts inventories, and reducing breakdown maintenance (Siddiqui, Sahay,

& Giri, 2014). Figure 1.1 presents three main maintenance strategies.

Figure 1.1 Three main maintenance strategies

4

1.2.1 Condition Monitoring for Fault Diagnosis Prediction

Condition monitoring leading to fault detection of IMs has been an attractive

research area in the last few years because of its significant effect in many industrial

processes. Correct detection and early prediction of incipient faults consequence in fast

unscheduled maintenance and short downtime for the process under consideration.

Destructive consequences can be avoided by condition monitoring. Financial loss also is

reduced. An ideal diagnostic technique should provide the minimum essential

measurements from a motor (Jin, Zhao, Chow, & Pecht, 2014; Toliyat, Nandi, Choi, &

Meshgin-Kelk, 2012).

In the scope of industry, most of the occurred faults are not predictable or even

visible with the naked eye. Therefore, it is very critical to identify and diagnose these

faults at early stages to prevent any corruption or damages in electrical instruments. For

example, since the air gap between rotor and stator is very small, any imbalance in barriers

or mis-positioning of rotor may cause serious physical damages to the rotor and stator of

the IM (Bellini, Filippetti, Tassoni, & Capolino, 2008).

Different monitoring procedures have been utilized for fault detection on IMs.

Vibration analysis, stray flux, and stator current-signature analysis (SCSA) are the most

popular ones (A Bellini, Concari, Franceschini, Tassoni, & Toscani, 2006). Stator faults

result in the open or short circuits on one or more stator windings (V Spyropoulos & D

Mitronikas, 2013). Extreme heating, transient over voltages, winding movement, or

contamination are the factors providing the winding-insulation damage. This fault causes

in high currents and winding overheating, which result in severe phase-to-phase, turn-to-

turn, or turn-to-ground faults. All these may lead to an irreversible damage in the windings

or in the stator core. Hence, affordable and reliable diagnosis of incipient faults between

turns during motor operation is vital (El Hachemi Benbouzid, 2000; Nandi, Toliyat, & Li,

2005; Tallam et al., 2007; Torkaman).

5

1.3 Methods for Fault Diagnosis of IMs

The existing methods for the fault diagnosis of IMs can be generally categorized

into three groups, namely: model-based, signal-based, and data-based (Alberto Bellini,

Filippetti, Tassoni, & Capolino, 2008). Most of the diagnostic techniques for IMs can be

extended easily to other types of rotating electrical machines.

1.3.1 Model-Based Fault Detection Method

Model-based fault detection method depends in light of a theoretical analysis of

the asymmetrical motor whose model is utilized to anticipate fault signatures (Alberto

Bellini et al., 2008; Isermann, 2005; Siddiqui et al., 2014). The difference between

measured and simulated signatures is used as a fault detector as shown in Figure 1.2.

Figure 1.2 Model-based diagnostic technique. Two techniques are possible according to

the same basic structure (Bellini et al., 2008)

6

In order to express a fault index, residual analysis and proper signal processing are

usually utilized (Bellini et al., 2008). Some left-overs are generated by model-based fault

detection and diagnosis methods which is indication of variations between measurement

and prediction. Theoretically, System faults only affect these left-over signals and the

deviations in the system inputs and predicted disturbances faced in normal operating

conditions have almost no effect on them. Power supply imbalances and load variations

are two critical parameters for electric motors. Hence, for normal condition and operating

without any faults the left-overs must be almost zero-mean white noise while in the case

of any faults they must deviate from this behavior. Model-based fault detection method

has not been prevalent and popular to be applied in the industry due to complications in

obtaining accurate and suitable models while modeling uncertainties resulting from

system nonlinearities, parameter uncertainties, disturbances and other measurement noise

exist (Combastel et al., 2002). Moreover, modeling of electromechanical systems is not

practical due to their complex construction and the requirement of extensive

approximations, which makes model-based analysis methods an inappropriate choice

(Combastel, Lesecq, Petropol, & Gentil, 2002; Kim & Parlos, 2002).

1.3.2 Signal-Based Fault Detection Method

Signal-based methods mostly focus on frequency domain data. The known fault

signatures in quantities sampled from the actual machine are detected by signal-based

diagnosis (Bellini et al., 2008). The signs are examined and observed by a proper signal

processing unit as shown in Figure 1.3. Even though advanced methods and/ or decision-

making techniques can be used, frequency analysis is normally used. In this method, signal

processing has an important role since it can improve signal-to-noise ratio and normalize

data to differentiate other faults generated from other sources. It is also able to reduce the

sensitivity to operating conditions (Bellini et al., 2008; Kim & Parlos, 2002). The signal-

based systems are mostly utilized for the procedures in the steady state. Effectiveness of

such fault diagnosis method in dynamic systems is significantly limited.

7

Figure 1.3 Block diagram of signal-based diagnostic procedure (Bellini et al., 2008)

1.3.3 Data-Based Fault Detection Method

Data-based diagnosis relies on signal processing and on classification methods.

The data-based techniques are considered more suitable options as a result of any

information of machine parameters and model is not required in this type of fault detection

technique (Bellini et al., 2008). In that perspective, such fault diagnosis offers a few

numbers of mathematical calculations. They are applied on the lines of the supervised

learning methods. In the supervised learning, the data are collected from the system in

known health conditions and based on the decision rules developed, health conditions of

unknown systems are categorized and prognosticated (Kim & Parlos, 2002).

The advantage of using data-based diagnosis is that it does not need any

information of machine model and parameters. Signal processing and clustering methods

8

are only requirement in this technique. Sample data are captured from an actual IM and

are processed in order to find a set features for classification purpose. Finally, fault index

can be achieved by utilizing decision process techniques as shown in Figure 1.4.

Figure 1.4 Block diagram of data-based diagnostic procedure (Bellini et al., 2008)

Data sampled from the motor are managed to extract a features' set that are

classified by classification methods. A fault index is defined by decision process

techniques. Artificial intelligence (AI) systems are broadly applied to classify faulty and

healthy conditions (Siddique, Yadava, & Singh, 2003).

9

1.4 AI Techniques for Motor Fault Diagnosis

In recent years, AI technologies have been employed to overcome the difficulties

that conventional diagnosis strategies (direct inspection, wear particle analysis and

parameter estimation) are facing. Conventional methods are easy to understand. However,

they are not always possible in reality because they require in-depth knowledge of the

induction motor system or its working mechanisms. In the case of inadequate information

false alarms can occur. A brief review of the advantages and disadvantages of these

approaches is given in Table 1.1 (Awadallah & Morcos, 2003; Gao & Ovaska, 2001;

Laghari, Memon, & Khuwaja, 2004; Tsang, 1995).

Table 1.1: Comparison of methods for motor fault detection and diagnosis

Approach Advantage Disadvantage

Direct inspection

Wear particle analysis

Parameter estimation

Expert systems (ES)

Artificial neural

network (ANN)

Simple & direct

Analysis theory is

mature and suitable for

routine check-up

Suitable for on-line

monitoring and fault

diagnosis

Known experience and

knowledge. Excellent

explanation capability

Without the need of

complex and rigorous

mathematical models or

expert experience

Requires experienced

engineers

Time consuming and

exhaustive

examination required

Difficult to obtain accurate

mathematical model

Expert experience &

knowledge is difficult to be

transformed & automated

Need training data

10

In general, expert systems and artificial neural networks (ANN) are one of the

most popular methods within AI systems. The effectiveness of the expert systems depends

on the precision and completeness of the knowledge base, which is usually very

complicated, time consuming and must be constructed manually. The major problem with

expert systems is that they cannot adjust their diagnostic rules automatically, and thus

cannot acquire knowledge from new data samples (Siddiqui et al., 2014).

ANN based method is rather easy to develop and perform. Unlike parameter

estimation and expert systems, ANN strategy can detect and diagnose motor faults based

on measurements without the need for complex and rigorous mathematical models or

experience. ANN systems can learn fault detection and diagnosis solely based on input-

output examples without the need of mathematical models. Therefore, ANN systems have

drawn significant attention in the motor fault detection and diagnosis field. No prior

knowledge about motor fault detection and diagnosis is needed. Only the training data,

including normal and faulty data need to be obtained in advance. Once ANNs are trained

appropriately, the networks could contain knowledge needed to perform fault detection

and diagnosis (Kumari & Sunita, 2013; Nasira, Kumar, & Kiruba, 2008; Shi, Sun, Li, &

Liu, 2007; Siddiqui et al., 2014).

1.5 Problem Statement

Electrical and mechanical data are commonly used in data-based diagnosis (Kano

& Nakagawa, 2008). The electrical current waveform of the IM can potentially reveal

whether the machine is working properly or not. It is notable that there are specific

characteristic behaviors in the current signals (provided by inverters) or vibration signals

(provided by accelerometer sensors placed on the machine) for each kind of main motor

faults. Therefore, it is feasible to detect the faults based on current and vibration

measurements (Garcia-Ramirez, Osornio-Rios, Granados-Lieberman, Garcia-Perez, &

Romero-Troncoso, 2012).

11

Motor current signature analysis (MCSA) is known as an effective technique for

fault diagnosis in three-phase IMs. This method is associated to various faults such as

broken rotor bars and windings faults. Numerous technical works have been recently

studied the benefits of this method in detecting the IM faults (El Hachemi Benbouzid,

2000; Penman et al., 1994; Radhika et al., 2010; Sadri, 2004; J.-W. Zhang et al., 2007; Z.

Zhang et al., 2003). Current sensors are mostly cheap and could be used and maintained

easily. Tandon et.al (2007) reported that stator current monitoring requires minimum

instruments and can be considered as an affordable fault detection technique (Tandon,

Yadava, & Ramakrishna, 2007). However, it has some limitations that reduce the

performance and accuracy of motor diagnosis. Bellini et.al (2008) proposed that stator

current monitoring is not a reliable fault detection system because the current signal

analysis is effective for the faults whose critical frequency rate is lower than the supply

frequency. The current signal can be utilized as a reliable approach only in dedicated

operating conditions (Bellini, Immovilli, Rubini, & Tassoni, 2008).

Vibration monitoring technique is a powerful approach for fault diagnosis in IMs.

It has been widely used due to its significant results. Fault diagnosis based on mechanical

features such as vibration of the stator furnishes the operator with high accuracy of results

(Dorrell, Thomson, & Roach, 1997). The dark side of such technique is the high cost of

accelerometers and associated wiring, which also require expensive software and

technical assistant to be utilized as reported by Nandi et.al (2005). They stated that

vibration transducers are expensive and require special installation conditions to avoid

harm owing to shock and vibration (Nandi, Toliyat, & Li, 2005). Thus, its use in several

applications may be limited. Subsequently, this method cannot utilize for large machines

fault diagnostics purpose because it is expensive (Siddiqui, Sahay, & Giri, 2014). The

vibration sensors could be damaged easily as well, which makes them improper for being

used in rough industrial environments (Gritli, Filippetti, Miceli, Rossi, & Chatti, 2012).

12

In a research done by Bellini et.al, use of vibration and current signal was

compared, in order to show advantages and disadvantages of this two condition

monitoring systems. They utilized the frequency domain to analyze capturing data. Signal

processing methods including Hilbert transformation and Envelope analysis were used for

machines with healthy and faulty bearings to demonstrate which monitoring system is the

best suited to the bearing failures. They found out that current signal cannot be considered

as a reliable fault detection system because of the current signal analysis is effective for

the faults whose critical frequency rate is lower than the supply frequency. Vibration

monitoring technique, however, showed that can be a reliable but expensive indicator for

bearing faults in low and high frequency. Though, vibration needs a structural model with

mass, damping and stiffness parameters. On the other hand, frequency domain analysis

requires different types of signal processing methods with complex mathematical

equations (Bellini, Immovilli, Rubini, & Tassoni, 2008).

Rodenas and Daviu proposed a twofold method for detection of broken rotor bars,

cooling system problems and bearing faults in IMs. The first stage utilized current

monitoring technique using steady state and transient methods. They used infrared

cameras to take thermography images to find failure places in a second stage. Although,

each of these approaches provided useful information to detect extensive ranges of faults,

but they were applicable for large and expensive motors. The infrared technique was

sensitive to failures located near the machine frame surface rather than to internal faults.

Furthermore, infrared cameras are so expensive. Another limitation was the length of the

required data due to the long duration of the heating transient. This system also may not

be applicable in industrial area with high temperature environment. Therefore, an

insensitive to heat and cost-effective fault diagnosis approach is required to be affordable

for all types of motors not only large and expensive ones. Besides, an ideal fault detection

technique should diagnose failures at inner and outer parts of machines (Picazo-Ródenas,

Antonino-Daviu, Climente-Alarcon, Royo-Pastor, & Mota-Villar, 2015).

13

There is almost no single fault diagnosis method capable to detect all probable

faults taking place in IMs with a reasonable price and high accuracy. Although stator

current and vibration monitoring are the most commonly used monitoring procedures in

the industry, but each of these techniques alone have some limitations. Consequently, a

single fault detection technique cannot be considered as a reliable and general diagnosis

system. While current monitoring technique is an inexpensive method, but it is less

accurate. Vibration monitoring on the other hand has higher price and accuracy compared

to the current monitoring. It must be noted that systems required high accuracy and lower

cost. Therefore, a new method for fault detection is deeply needed to meet these

requirements. This thesis presented a cost-effective and reliable method for detection of

faults in three-phase IMs by combination of the two aforementioned monitoring

techniques (vibration and current) with great prospect for application in industrial scale.

1.6 Objectives of the Study

The objectives of this thesis are as follows:

(1) To develop an affordable installation and maintenance setup for fault

diagnosis in IMs.

(2) To develop an intelligent fault detection strategy based on vibration and

electrical current signals.

(3) To evaluate the performance of vibration and current setup in term of

accuracy and cost.

1.7 Scope of the Study

This investigation was conducted to determine the stator winding and broken rotor

bar faults in three phase induction machines with a squirrel cage rotor. Two faulty IMs

with broken rotor and winding faults and one healthy IM have been investigated in the

14

Center for Artificial Intelligent and Robotics (CAIRO) laboratory at University Teknologi

Malaysia (UTM). Data were captured by two different setups in time domain:

i) Vibration setup contains NI PCI- 4474 DAQ card and accelerometers

ii) Current setup included NI 9234 and NI 9174 CDAQ cards, and current clipping

sensor.

Each of these two setups alone have some limitations for fault diagnosis in IMs.

Vibration setup is expensive, whilst current setup is cheap but with low detection

reliability. This research work assumes to develop a reliable and cost-effective fault

detection method with the joint use of vibration and current setups. In addition, ANN was

used for classification and nonlinear regression system. PCA technique also utilized for

reduction of features dimensions.

1.8 Thesis Organization

This thesis is organized into five chapters. A brief outline of the thesis’s

contents is as follows:

Chapter 1 presents an introduction to the research problem. It involves the

background of the study, problem statement and hypothesis of the thesis. The logical

flow and structure of the thesis are also outlined in this chapter.

A complete literature review on faulty IMs with various types of faults,

condition monitoring techniques, different methods for fault detection and their

advantages and disadvantages are presented in chapter 2.

Chapter 3 focuses on the proposed methodology contained data acquisition,

feature extraction, method development including testing set to train the algorithm and

REFERENCES

Arthur, N., & Penman, J. (2000). Induction machine condition monitoring with higher

order spectra. Industrial Electronics, IEEE Transactions on, 47(5), 1031-1041.

Awadallah, M., & Morcos, M. M. (2003). Application of AI tools in fault diagnosis of

electrical machines and drives-an overview. Energy Conversion, IEEE

Transactions on, 18(2), 245-251.

Ayaz, E. (2014). A Review Study on Mathematical Methods for Fault Detection Problems

in Induction Motors. Balkan Journal of Electrical and Computer Engineering,

2(3).

Beguenane, R., & Benbouzid, M. E. H. (2000). Induction motors thermal monitoring by

means of rotor resistance identification. IEEE transactions on energy conversion,

14(3), 566-570.

Bellini, A., Concari, C., Franceschini, G., Tassoni, C., & Toscani, A. (2006). Vibrations,

currents and stray flux signals to asses induction motors rotor conditions. Paper

presented at the IEEE Industrial Electronics, IECON 2006-32nd Annual

Conference on.

Bellini, A., Filippetti, F., Franceschini, G., Tassoni, C., & Kliman, G. B. (2001).

Quantitative evaluation of induction motor broken bars by means of electrical

signature analysis. Industry Applications, IEEE Transactions on, 37(5), 1248-

1255.

Bellini, A., Filippetti, F., Tassoni, C., & Capolino, G.-A. (2008). Advances in diagnostic

techniques for induction machines. IEEE Transactions on Industrial Electronics,

12(55), 4109-4126.

Blödt, M., Chabert, M., Regnier, J., & Faucher, J. (2006). Mechanical load fault detection

in induction motors by stator current time-frequency analysis. Industry

Applications, IEEE Transactions on, 42(6), 1454-1463.

92

Borras, D., Castilla, M., Moreno, N., & Montano, J. (2001). Wavelet and neural structure:

a new tool for diagnostic of power system disturbances. Industry Applications,

IEEE Transactions on, 37(1), 184-190.

Butler, K. L. (1996). An expert system based framework for an incipient failure detection

and predictive maintenance system. Paper presented at the Intelligent Systems

Applications to Power Systems, 1996. Proceedings, ISAP'96., International

Conference on.

Cabanas, M., Melero, M., Orcajo, G., Faya, F. R., & Sariego, J. S. (1998). Experimental

application of axial leakage flux to the detection of rotor asymmetries, mechanical

anomalies and interturn short circuits in working induction motors. Paper

presented at the Proc ICEM'98.

Cao, L., Chua, K., Chong, W., Lee, H., & Gu, Q. (2003). A comparison of PCA, KPCA

and ICA for dimensionality reduction in support vector machine. Neurocomputing,

55(1), 321-336.

Combastel, C., Lesecq, S., Petropol, S., & Gentil, S. (2002). Model-based and wavelet

approaches to induction motor on-line fault detection. Control Engineering

Practice, 10(5), 493-509.

Corres, J. M., Bravo, J., Arregui, F. J., & Matias, I. R. (2006). Vibration monitoring in

electrical engines using an in-line fiber etalon. Sensors and Actuators A: Physical,

132(2), 506-515.

Cusidó, J., Romeral, L., Ortega, J. A., Garcia, A., & Riba, J. (2011). Signal injection as a

fault detection technique. Sensors, 11(3), 3356-3380.

Da Silva, A. M. (2006). Induction motor fault diagnostic and monitoring methods. Faculty

Of the Graduate School, Marquette University.

Da Silva, A. M., Povinelli, R. J., & Demerdash, N. A. (2013). Rotor bar fault monitoring

method based on analysis of air-gap torques of induction motors. IEEE

Transactions on Industrial Informatics, 9(4), 2274-2283.

Delgado Prieto, M., Cusido Roura, J., & Romeral Martínez, J. L. (2011). Bearings fault

detection using inference tools: In-Tech.

Dorrell, D. G., Thomson, W. T., & Roach, S. (1997). Analysis of airgap flux, current, and

vibration signals as a function of the combination of static and dynamic airgap

93

eccentricity in 3-phase induction motors. Industry Applications, IEEE

Transactions on, 33(1), 24-34.

Douglas, H., Pillay, P., & Ziarani, A. K. (2004). A new algorithm for transient motor

current signature analysis using wavelets. Industry Applications, IEEE

Transactions on, 40(5), 1361-1368.

Dowding, C. H. (1985). Blast vibration monitoring and control (Vol. 297): Prentice-Hall

Englewood Cliffs.

Drucker, H., Wu, D., & Vapnik, V. N. (1999). Support vector machines for spam

categorization. Neural Networks, IEEE Transactions on, 10(5), 1048-1054.

Dutta, H., Giannella, C., Borne, K. D., & Kargupta, H. (2007). Distributed Top-K Outlier

Detection from Astronomy Catalogs using the DEMAC System. Paper presented at

the SDM.

Ekici, S., Yildirim, S., & Poyraz, M. (2008). Energy and entropy-based feature extraction

for locating fault on transmission lines by using neural network and wavelet packet

decomposition. Expert Systems with Applications, 34(4), 2937-2944.

El Hachemi Benbouzid, M. (2000). A review of induction motors signature analysis as a

medium for faults detection. Industrial Electronics, IEEE Transactions on, 47(5),

984-993.

Ellison, A., & Yang, S. (1971). Effects of rotor eccentricity on acoustic noise from

induction machines. Paper presented at the Proceedings of the Institution of

Electrical Engineers.

Filippetti, F., Franceschini, G., Tassoni, C., & Vas, P. (1998). AI techniques in induction

machines diagnosis including the speed ripple effect. Industry Applications, IEEE

Transactions on, 34(1), 98-108.

Filippetti, F., Franceschini, G., Tassoni, C., & Vas, P. (2000). Recent developments of

induction motor drives fault diagnosis using AI techniques. Industrial Electronics,

IEEE Transactions on, 47(5), 994-1004.

Filippetti, F., Martelli, M., Franceschini, G., & Tassoni, C. (1992). Development of expert

system knowledge base to on-line diagnosis of rotor electrical faults of induction

motors. Paper presented at the Industry Applications Society Annual Meeting,

1992., Conference Record of the 1992 IEEE.

94

Finley, W. R., Hodowanec, M. M., & Holter, W. G. (1999). An analytical approach to

solving motor vibration problems. Paper presented at the Petroleum and Chemical

Industry Conference, 1999. Industry Applications Society 46th Annual.

Fujimaki, R., Yairi, T., & Machida, K. (2005). An approach to spacecraft anomaly

detection problem using kernel feature space. Paper presented at the Proceedings

of the eleventh ACM SIGKDD international conference on Knowledge discovery

in data mining.

Fukuyama, Y., & Ueki, Y. (1992). Hybrid fault analysis system using neural networks

and artificial intelligence. Paper presented at the Circuits and Systems, 1992.

ISCAS'92. Proceedings., 1992 IEEE International Symposium on.

Gao, X. Z., & Ovaska, S. J. (2001). Soft computing methods in motor fault diagnosis.

Applied soft computing, 1(1), 73-81.

Garcia-Ramirez, A. G., Osornio-Rios, R. A., Granados-Lieberman, D., Garcia-Perez, A.,

& Romero-Troncoso, R. J. (2012). Smart sensor for online detection of multiple-

combined faults in VSD-fed induction motors. Sensors, 12(9), 11989-12005.

Ghate, V. N., & Dudul, S. V. (2009). Fault diagnosis of three phase induction motor using

neural network techniques. Paper presented at the Emerging Trends in

Engineering and Technology (ICETET), 2009 2nd International Conference on.

Godoy, W. F., da Silva, I. N., Goedtel, A., Palácios, R. H., Bazan, G. H., & Moríñigo-

Sotelo, D. (2016). An application of artificial neural networks and PCA for stator fault

diagnosis in inverter-fed induction motors. Paper presented at the Electrical Machines

(ICEM), 2016 XXII International Conference on.

Goode, P. V., & Chow, M.-Y. (1994). A hybrid fuzzy/neural system used to extract

heuristic knowledge from a fault detection problem. Paper presented at the Fuzzy

Systems, 1994. IEEE World Congress on Computational Intelligence.,

Proceedings of the Third IEEE Conference on.

Gritli, Y., Filippetti, F., Miceli, R., Rossi, C., & Chatti, A. (2012). Investigation of motor

current signature and vibration analysis for diagnosing rotor broken bars in

double cage induction motors. Paper presented at the Power Electronics, Electrical

Drives, Automation and Motion (SPEEDAM), 2012 International Symposium on.

95

Günter, S., Schraudolph, N., & Vishwanathan, S. (2007). Fast iterative kernel principal

component analysis. Made available in DSpace on 2010-12-20T06: 02: 43Z

(GMT). No. of bitstreams: 1 Guenter_Fast2007. pdf: 3186111 bytes, checksum:

068a6868a41936f686cc8c5e9412eaaa (MD5) Previous issue date: 2009-05-

21T05: 19: 22Z.

Han, T., Yang, B.-S., Choi, W.-H., & Kim, J.-S. (2006). Fault diagnosis system of

induction motors based on neural network and genetic algorithm using stator

current signals. International Journal of Rotating Machinery, 2006.

Hawkins, D. M. (1974). The detection of errors in multivariate data using principal

components. Journal of the American Statistical Association, 69(346), 340-344.

Horvath, G. (2003). CMAC neural network as an SVM with B-Spline kernel functions.

Paper presented at the Instrumentation and Measurement Technology Conference,

2003. IMTC'03. Proceedings of the 20th IEEE.

Huang, Z., Chen, H., Hsu, C.-J., Chen, W.-H., & Wu, S. (2004). Credit rating analysis

with support vector machines and neural networks: a market comparative study.

Decision support systems, 37(4), 543-558.

Hsu, J. S. (1995). Monitoring of defects in induction motors through air-gap torque

observation. Industry Applications, IEEE Transactions on, 31(5), 1016-1021.

Ide, T., & Kashima, H. (2004). Eigenspace-based anomaly detection in computer systems.

Paper presented at the Proceedings of the tenth ACM SIGKDD international

conference on Knowledge discovery and data mining.

Isermann, R. (2005). Model-based fault-detection and diagnosis–status and applications.

Annual Reviews in control, 29(1), 71-85.

Jardine, A. K., Lin, D., & Banjevic, D. (2006). A review on machinery diagnostics and

prognostics implementing condition-based maintenance. Mechanical systems and

signal processing, 20(7), 1483-1510.

Jia, F., Lei, Y., Lin, J., Zhou, X., & Lu, N. (2016). Deep neural networks: A promising

tool for fault characteristic mining and intelligent diagnosis of rotating machinery

with massive data. Mechanical systems and signal processing, 72, 303-315.

96

Jin, X., Zhao, M., Chow, T. W., & Pecht, M. (2014). Motor bearing fault diagnosis using

trace ratio linear discriminant analysis. Industrial Electronics, IEEE Transactions

on, 61(5), 2441-2451.

Jolliffe, I. (2002). Principal component analysis: Wiley Online Library.

Kano, M., & Nakagawa, Y. (2008). Data-based process monitoring, process control, and

quality improvement: Recent developments and applications in steel industry.

Computers & Chemical Engineering, 32(1), 12-24.

Kim, K., & Parlos, A. G. (2002). Induction motor fault diagnosis based on neuropredictors

and wavelet signal processing. Mechatronics, IEEE/ASME Transactions on, 7(2),

201-219.

Kumari, N., & Sunita, S. (2013). Comparison of ANNs, Fuzzy Logic and Neuro-Fuzzy

Integrated Approach for Diagnosis of Coronary Heart Disease: A Survey.

Kwitowski, A. J., Lewis, W. H., & Mayercheck, W. D. (1989). Computer-based,

teleoperation of a new highwall mining system. Paper presented at the Robotics

and Automation, 1989. Proceedings., 1989 IEEE International Conference on.

Laghari, M. S., Memon, Q. A., & Khuwaja, G. A. (2004). Knowledge based wear particle

analysis. International Journal of Information Technology, 1(3), 91-95.

Lakhina, A., Crovella, M., & Diot, C. (2005). Mining anomalies using traffic feature

distributions. Paper presented at the ACM SIGCOMM Computer Communication

Review.

Luo, Y., Zhixiang, X., Sun, H., Yuan, S., & Yuan, J. (2015). Research on the induction

motor current signature for centrifugal pump at cavitation condition. Advances in

Mechanical Engineering, 7(11), 1687814015617134.

Malhotra, S., & Soni, M. (2013). Fault diagnosis of induction motor. European Scientific

Journal, 9(21).

Martins, J. F., Pires, V. F., & Pires, A. (2007). Unsupervised neural-network-based

algorithm for an on-line diagnosis of three-phase induction motor stator fault.

IEEE transactions on industrial electronics, 54(1), 259-264.

Krizhevsky, A., Sutskever, I., & Hinton, G. E. (2012). Imagenet classification with deep

convolutional neural networks. Paper presented at the Advances in neural

information processing systems.

97

Luo, Y., Zhixiang, X., Sun, H., Yuan, S., & Yuan, J. (2015). Research on the induction

motor current signature for centrifugal pump at cavitation condition. Advances in

Mechanical Engineering, 7(11), 1687814015617134.

Masnadi-Shirazi, H., & Vasconcelos, N. (2009). On the design of loss functions for

classification: theory, robustness to outliers, and savageboost. Paper presented at

the Advances in neural information processing systems.

Radford, A., Metz, L., & Chintala, S. (2015). Unsupervised representation learning with

deep convolutional generative adversarial networks. arXiv preprint

arXiv:1511.06434.

Vico, J., Stankovic, D., Voloh, I., Zhang, Z., & Swigost, D. (2010). Enhanced algorithm

for motor rotor broken bar detection.

Vitek, O., Janda, M., Hajek, V., & Bauer, P. (2011). Detection of eccentricity and bearings

fault using stray flux monitoring. Paper presented at the Diagnostics for Electric

Machines, Power Electronics & Drives (SDEMPED), 2011 IEEE International

Symposium on.

Wan, L., Zeiler, M., Zhang, S., Le Cun, Y., & Fergus, R. (2013). Regularization of neural

networks using dropconnect. Paper presented at the International Conference on

Machine Learning.

McCormick, A. C. (1998). Cyclostationary and higher-order statistical signal processing

algorithms for machine condition monitoring. Citeseer.

McFadden, P., & Smith, J. (1984). Vibration monitoring of rolling element bearings by

the high-frequency resonance technique—a review. Tribology international,

17(1), 3-10.

Mehala, N. (2010). Condition monitoring and fault diagnosis of induction motor using

motor current signature analysis. National institute of technology kurukshetra,

India.

Mellor, P., Roberts, D., & Turner, D. (1991). Lumped parameter thermal model for

electrical machines of TEFC design. Paper presented at the IEE Proceedings B

(Electric Power Applications).

98

Milanfar, P., & Lang, J. H. (1996). Monitoring the thermal condition of permanent-magnet

synchronous motors. Aerospace and Electronic Systems, IEEE Transactions on,

32(4), 1421-1429.

Mohri, M., Rostamizadeh, A., & Talwalkar, A. (2012). Foundations of machine learning:

MIT press.

Moreno, J. F., Hidalgo, F. P., & Martinez, M. (2001). Realisation of tests to determine the

parameters of the thermal model of an induction machine. Paper presented at the

Electric Power Applications, IEE Proceedings.

Nasira, G., Kumar, S. A., & Kiruba, M. S. (2008). A Comparative Study Of Fuzzy Logic

With Artificial Neural Networks Algorithms In Clustering. Journal of Computer

Applications, 1(4), 6.

Okoro, O. (2005). Steady and transient states thermal analysis of a 7.5-kW squirrel-cage

induction machine at rated-load operation. Energy Conversion, IEEE Transactions

on, 20(4), 730-736.

Palem, G. (2013). Condition-based maintenance using sensor arrays and telematics. arXiv

preprint arXiv:1309.1921.

Paoletti, G., & Golubev, A. (1999). Partial discharge theory and applications to electrical

systems. Paper presented at the Pulp and Paper, 1999. Industry Technical

Conference Record of 1999 Annual.

Parra, L., Deco, G., & Miesbach, S. (1996). Statistical independence and novelty detection

with information preserving nonlinear maps. Neural Computation, 8(2), 260-269.

Payne, D., Ball, A., & Gu, F. (2002). Detection and diagnosis of induction motor faults

using statistical measures. International Journal of COMADEM, 5(2), 5-19.

Penman, J., Sedding, H., Lloyd, B., & Fink, W. (1994). Detection and location of interturn

short circuits in the stator windings of operating motors. Energy Conversion, IEEE

Transactions on, 9(4), 652-658.

Pham, D., & Pham, P. (1999). Artificial intelligence in engineering. International Journal

of Machine Tools and Manufacture, 39(6), 937-949.

Picazo-Ródenas, M. J., Antonino-Daviu, J., Climente-Alarcon, V., Royo-Pastor, R., &

Mota-Villar, A. (2015). Combination of noninvasive approaches for general

99

assessment of induction motors. IEEE Transactions on Industry Applications,

51(3), 2172-2180.

Premrudeepreechacharn, S., Utthiyoung, T., Kruepengkul, K., & Puongkaew, P. (2002).

Induction motor fault detection and diagnosis using supervised and unsupervised

neural networks. Paper presented at the Industrial Technology, 2002. IEEE

ICIT'02. 2002 IEEE International Conference on.

Radhika, S., Sabareesh, G., Jagadanand, G., & Sugumaran, V. (2010). Precise wavelet for

current signature in 3ϕ IM. Expert Systems with Applications, 37(1), 450-455.

Riley, C. M., Lin, B. K., Habetler, T. G., & Kliman, G. B. (1997). Stator current-based

sensorless vibration monitoring of induction motors. Paper presented at the

Applied Power Electronics Conference and Exposition, 1997. APEC'97

Conference Proceedings 1997., Twelfth Annual.

Riley, C. M., Lin, B. K., Habetler, T. G., & Kliman, G. B. (1999). Stator current harmonics

and their causal vibrations: a preliminary investigation of sensorless vibration

monitoring applications. Industry Applications, IEEE Transactions on, 35(1), 94-

99.

Sadri, H. (2004). Induction Motor Broken Rotor Bar Fault Detection Using Wavelet

Analysis. Master's Thesis, Iran University of Science & Technology.

Shi, L., Sun, Z., Li, H., & Liu, W. (2007). Research on diagnosing coronary heart disease

using fuzzy adaptive resonance theory mapping neural networks. Paper presented

at the Control and Automation, 2007. ICCA 2007. IEEE International Conference

on.

Shyu, M.-L., Chen, S.-C., Sarinnapakorn, K., & Chang, L. (2003). A novel anomaly

detection scheme based on principal component classifier: DTIC Document.

Siddique, A., Yadava, G., & Singh, B. (2003). Applications of artificial intelligence

techniques for induction machine stator fault diagnostics: review. Paper presented

at the Diagnostics for Electric Machines, Power Electronics and Drives, 2003.

SDEMPED 2003. 4th IEEE International Symposium on.

Siddiqui, K. M., & Giri, V. (2012). Broken rotor bar fault detection in induction motors

using wavelet transform. Paper presented at the Computing, Electronics and

Electrical Technologies (ICCEET), 2012 International Conference on.

100

Siddiqui, K. M., Sahay, K., & Giri, V. (2014). Health Monitoring and Fault Diagnosis in

Induction Motor-A Review. International Journal of Advanced Research in

Electrical, Electronics and Instrumentation Engineering, 3(1), 6549-6565.

Stikeleather, L. F. (1976). Review of ride vibration standards and tolerance criteria: SAE

Technical Paper.

Styvaktakis, E., Bollen, M. H., & Gu, I. Y. (2002). Expert system for classification and

analysis of power system events. Power Delivery, IEEE Transactions on, 17(2),

423-428.

Sun, J., Xie, Y., Zhang, H., & Faloutsos, C. (2007). Less is more: Compact matrix

representation of large sparse graphs. Paper presented at the Proceedings of 7th

SIAM International Conference on Data Mining.

Tandon, N., Yadava, G., & Ramakrishna, K. (2007). A comparison of some condition

monitoring techniques for the detection of defect in induction motor ball bearings.

Mechanical systems and signal processing, 21(1), 244-256.

Thomson, W. T., & Fenger, M. (2001). Current signature analysis to detect induction

motor faults. Industry Applications Magazine, IEEE, 7(4), 26-34.

Thottan, M., & Ji, C. (2003). Anomaly detection in IP networks. Signal Processing, IEEE

Transactions on, 51(8), 2191-2204.

Toliyat, H. A., Nandi, S., Choi, S., & Meshgin-Kelk, H. (2012). Electric Machines:

Modeling, Condition Monitoring, and Fault Diagnosis: CRC Press.

Torkaman, H., Afje, E., & Yadegari, P. (2012). Static, dynamic, and mixed eccentricity

faults diagnosis in switched reluctance motors using transient finite element

method and experiments. Magnetics, IEEE Transactions on, 48(8), 2254-2264.

Trutt, F. C., Sottile, J., & Kohler, J. L. (2002). Online condition monitoring of induction

motors. Industry Applications, IEEE Transactions on, 38(6), 1627-1632.

Tsang, A. H. (1995). Condition-based maintenance: tools and decision making. Journal

of Quality in Maintenance Engineering, 1(3), 3-17.

V Spyropoulos, D., & D Mitronikas, E. (2013). A Review on the Faults of Electric

Machines Used in Electric Ships. Advances in Power Electronics, 2013.

Valyon, J., & Horváth, G. (2003). A weighted generalized ls-svm. Periodica Polytechnica

Electrical Engineering, 47(3-4), 229-252.

101

Vas, P. (1993). Parameter estimation, condition monitoring, and diagnosis of electrical

machines: Oxford University Press, USA.

Vas, P. (1999). Artificial-intelligence-based electrical machines and drives: application

of fuzzy, neural, fuzzy-neural, and genetic-algorithm-based techniques (Vol. 45):

Oxford University Press.

Vico, J., Stankovic, D., Voloh, I., Zhang, Z., & Swigost, D. (2010). Enhanced algorithm

for motor rotor broken bar detection.

Vitek, O., Janda, M., Hajek, V., & Bauer, P. (2011). Detection of eccentricity and bearings

fault using stray flux monitoring. Paper presented at the Diagnostics for Electric

Machines, Power Electronics & Drives (SDEMPED), 2011 IEEE International

Symposium on.

Wang, X., & Elbuluk, M. (1996). Neural network control of induction machines using

genetic algorithm training. Paper presented at the Industry Applications

Conference, 1996. Thirty-First IAS Annual Meeting, IAS'96., Conference Record

of the 1996 IEEE.

Widodo, A., & Yang, B.-S. (2007). Application of nonlinear feature extraction and

support vector machines for fault diagnosis of induction motors. Expert Systems

with Applications, 33(1), 241-250.

Widodo, A., Yang, B.-S., & Han, T. (2007). Combination of independent component

analysis and support vector machines for intelligent faults diagnosis of induction

motors. Expert Systems with Applications, 32(2), 299-312.

Yam, R., Tse, P., Li, L., & Tu, P. (2001). Intelligent predictive decision support system

for condition-based maintenance. The International Journal of Advanced

Manufacturing Technology, 17(5), 383-391.

Yamamura, S. (1979). Theory of linear induction motors. New York, Halsted Press, 1979.

246 p., 1.

Yang, H., Mathew, J., & Ma, L. (2003). Vibration feature extraction techniques for fault

diagnosis of rotating machinery: a literature survey.

Ye, Z., Wu, B., & Sadeghian, A. (2003). Current signature analysis of induction motor

mechanical faults by wavelet packet decomposition. Industrial Electronics, IEEE

Transactions on, 50(6), 1217-1228.

102

Zarei, J. (2012). Induction motors bearing fault detection using pattern recognition

techniques. Expert Systems with Applications, 39(1), 68-73.

Zhang, J.-W., Zhu, N.-H., Li, Y., Qi, Y., & Qing, L. (2007). A fault diagnosis approach

for broken rotor bars based on EMD and envelope analysis. Journal of China

University of Mining and Technology, 17(2), 205-209.

Zhang, J. S. Y. X. H., & Faloutsos, C. (2007). Less is more: Compact matrix

decomposition for large sparse graphs. Paper presented at the Proceedings of the

Seventh SIAM International Conference on Data Mining.

Zhang, P., Du, Y., Habetler, T. G., & Lu, B. (2011). A survey of condition monitoring and

protection methods for medium-voltage induction motors. Industry Applications,

IEEE Transactions on, 47(1), 34-46.

Zhang, Z., Ren, Z., & Huang, W. (2003). A novel detection method of motor broken rotor

bars based on wavelet ridge. Energy Conversion, IEEE Transactions on, 18(3),

417-423.

Zolfaghari, S., Noor, M., Bahari, S., Mariun, N., Marhaban, M. H., Mehrjou, M. R., &

Karami, M. (2014). Broken rotor bar detection of induction machine using wavelet

packet coefficient-related features. Paper presented at the Research and

Development (SCOReD), 2014 IEEE Student Conference on.

thesisStatusDeclaration (new).pdf (p.1)R.Yousefi, PHD thesis, Aug 2018.pdf (p.2-3)R.Yousefi, PHD thesis, Nov 2018, printing.pdf (p.4-151)